Sticky Notes: Some Thoughts on the Use of Clay in the Neolithic Deposits within the Avebury Megalithic Complex

Studies of megalithic stone settings have tended to focus upon features perceptible above the ground, whereas sub-surface elements have received markedly less attention. Consideration of both as elements of assemblages provides opportunities for new understandings to emerge. Through re-examination of stone settings within the context of the Avebury monumental complex, the value of this approach is demonstrated. The existence of discrete material types with essential properties is undermined as new, fluid, contextually situated substances reveal themselves, enabling a reappraisal of other assemblages within the group.


INTRODUCTION
Encounters with megalithic monuments are corporeal and multi-sensory. As visually imposing, textural masses, it may be unsurprising that studies seeking to investigate the meaning of Neolithic megalithic monuments have traditionally focused upon extant stones, upon what is, or would once have been, visible above the ground surface. For individual stones, primary concerns have included size, shape, weight, colour, form, texture, source, biography and ontological status (Gillings and Pollard 1999, Tilley 2004, Parker Pearson 2012. Where grouped together, interest has centred upon layout with a particular emphasis on movement around and through alignments, as markers of space and associated concepts of inclusion and exclusion, as well as combinations of qualities of individual stones (Colt Hoare 1819, Bradley 1998, Cummings 2002, Tilley 2004, Lawson 2007, Gillings et al. 2008, Darvill and Wainwright 2009, Stukeley 2010, Parker Pearson 2012. Sub-surface features have received markedly less attention; where undertaken, examination of the significance of stone settings has tended to emphasise utilitarian concerns with the mechanics of stone erection and movement. There are notable exceptions (Cooney 2009, Gillings 2015. In his critical reappraisal of the role and meaning of the miniliths of Exmoor, of which their distinctive settings, including packing stones, are integral elements, Gillings (2015) offers an approach that provides new ways to engage with and understand these and other monuments. The importance of sub-surface features is recognised, thereby querying implicit hierarchies of visual and haptic dominance.
This article examines a group of megalithic settings within Avebury henge. Extant and sub-surface features of stone settings are considered together as assemblages of diverse substances. The relational materiality of these substances forms the focus of discussion, drawing upon the emergent qualities of the constituent materials, their sources and their subsequent redeployment, enabling a reconsideration of the nature and roles of the settings as ontologically articulate assemblages (Conneller 2011). Citation of material references (Jones 2012) is explored through extension of discussion to small pit deposits on the West Kennet occupation site, before the wider implications of such an approach are considered.

MEGALITHS, THE MARMALADE MAN AND A STICKY PROBLEM
Avebury megalithic complex is situated in a gently sloping valley, at the foot of Waden Hill in the north Wiltshire downs of southern England (Fig. 1). Constructed on chalk bedrock, it comprises an outer stone circle set within a substantial circular bank and ditch and two inner circles, one in the northern half and one in the southern half, alongside a number of individual stones. The northern inner circle surrounds the three standing stones of Avebury Cove, and within the southern circle there is evidence for the destruction of a stone Fig. 1. Key sites in the Avebury monumental complex mentioned within the text. Adapted from Gillings and Pollard (forthcoming, Fig. 1). With the kind permission of Mark Gillings and Joshua Pollard. named the Obelisk, described by Stukeley (2010, p. 24) as standing 21-feet tall, together with a linear arrangement of smaller stone settings. The henge has four causeways: the western and southern causeways are each conjoined with avenues of paired stones that extend across the adjacent low-lying land. The Beckhampton Avenue is over 1.3 km in length and links the western causeway with Longstones Cove, a multi-phase, megalithic setting that would have formed an open boxlike structure in its second incarnation (Gillings et al. 2008, pp. 124-126). The West Kennet Avenue leaves the southern causeway of Avebury Henge and tracks the base of Waden Hill, encountering the West Kennet occupation site and passing within view of Faulkner's Circle, another megalithic setting. It moves on to pass the West Kennet palisade enclosures, before joining with the Sanctuary on Overton Hill, covering a distance in excess of 2.4 km (Gillings et al. 2008, p. 129). Positioned within a dynamic landscape of construction and cultivation from the earliest Neolithic, evidenced by the presence of a dense concentration of long barrows, and middle Neolithic occupation on the West Kennet Avenue, the erection of the megalithic elements of Avebury complex spanned the 3rd millennium BC (Gillings et al. 2008, pp. 202-204).
Archaeological investigation of the Avebury megalithic settings has a long history, the structure and spatial configuration of the stones and earthworks emerging as dominant concerns from the outset, typified by accounts that privilege finished form over the process of construction (Colt Hoare 1819, Stukeley 2010Stukeley [1743). It seems telling that the meticulous excavation and recording of archaeological investigations undertaken in the 1930s at Avebury by Alexander Keiller, archaeologist and heir to the Keiller marmalade company fortune, was organised alongside a programme of restoration, whereby a number of post-medieval dwellings within the stone circle were removed, banks were levelled, buried stones were erected and, where stones were presumed to be absent, stone holes were marked with small concrete posts (Burl 1979, pp. 69-74, Pollard and Reynolds 2002, pp. 88-89, Gillings and Pollard 2004. An overarching concern with the visual impact of the monument as a cohesive architectural wholeand arguably, a failure to respond to the importance of sub-surface features as integral elements in all but a few instancesis demonstrated through Keiller's decision to reset the standing stones in substantial rafts of concrete (Fig. 2). Stone settings were, however, excavated carefully and the form and composition of fills documented (Fig. 3). Stone described as 'packing material' is noted in many of Avebury's megalithic settings, and in her account of material found by Keiller in the stone holes Smith (1965, p. 221) records the presence of smaller sarsen blocks and boulders. These stones are understood as a homogeneous class of material, a purely mechanistic device to provide structural stability: through use of the descriptor 'packing', a secondary (literally) supporting role is established. Smith also notes the inclusion of clean clay in deposits, again offering a utilitarian explanation for its presence: The most plausible explanation for the use of a plastic substance in the stone hole seems to be that it would have served to indicate whether or not the stone was properly balanced before the supporting timbers were removed. If the stone began to lean to one side a gap would appear between its base and the clay on the other. (1965, p. 221) This explanation seems unconvincing, Smith's choice of words revealing her uncertainty: the clay presents a problem that requires explaining away. Conceptualised as a discrete substance, a material type, lacking the properties necessary to afford direct, physical support in maintaining the megalith in an upright position, it is understood as malleable, but with the capacity to maintain form in the absence of forces with sufficient strength to cause change. The solution proffered feels distinctly modern, an architectural device perverse in its exploitation of material attributes that would appear contrary to the desired outcome. Also, this does not address the specified cleanness of the clay in deposits. Smith struggles to account adequately for the presence of the clay based on its mechanistic properties alone precisely because they are the product of modern, universalising, scientific engagements that fail to respond to contextual particularity. The interpretation of the small sarsens as straightforward packing material is thereby also undermined.

DYNAMISM, DELEUZE, DEPOSITION
The nature of materials as presented in Smith's account is both problematicdrawing on a set of implicit, untheorised, historically situated ontological assumptionsand pervasive, since this perspective underpins core archaeological principles, from the classification of periods, through to excavation practices and post-excavation analytical procedures, academic institutional structures and archaeological discourse (Conneller 2011, pp. 1-2). The scale of this issue and the consequences of attending to it seriously therefore have profound implications. Material categories serve as classificatory apparatus by which experiential existence may be explored and negotiated, and herein lies the catch. Barad (2007) problematises the nature of phenomena, with particular focus on scientific enquiry and the nature, role and boundedness of apparatus, rooted in a relational understanding of existence and the science philosophy of the pioneer of quantum theory, Niels Bohr. She argues that the apparatus employed in scientific study are implicated in the results obtained, allowing particular qualities to emerge whilst precluding others: the mode of questioning is constituent of and inseparable from the outcome. When thought about in this way, the adoption of a specific suite of material categories, whether conscious or otherwise, informs and places limitations upon the kinds of material properties and engagements that can present through analysis, and, for archaeological material, risks the projection of understandings that may shape and constrain interpretation.
Recent studies that draw on the assemblage theory of Deleuze and Guattari have much to offer here. Foregrounding associations and connections between phenomena, they understand materials not as static substances with a fixed suite of predetermined physical properties, but as vibrant assemblages with transient stability. Bennett (2010) examines the nature of matter, problematising simplistic, linear models of causation and arguing for diffused agency and dynamic materials that have the capacity to act and affect action, bound up in complex relationships that extend across time and space. These ideas are developed by Harris (2014) and Lucas (2012, pp. 169-214) in their explorations of archaeological assemblages, multi-scalar phenomena that destabilise traditional definitions of discrete materials, understood as both defined entities and events in process. They are the product of, and actively transform and create, diverse relationships through ongoing engagement, drawing together people associated with their creation and use, other materials, events and places (Harris 2014, pp. 331-332). Materials are thus vital, agentive and relationally constituted, and as such the existence of essential properties latent within them and recognised as such at all times and in all placesthe position implicit in Smith's interpretationis called into question.
This issue is a core concern of Conneller's (2011) volume exploring the role and meaning of archaeological materials, in which she argues persuasively that material qualities are relational, emergent through association and practice. Although focused primarily on consideration of material from the Upper Palaeolithic and Mesolithic, her approach places emphasis on the importance of attendance to the materials themselves, artefacts examined as materials, assemblages, forms and techniques, through relations, providing a useful tool for a reconsideration of the role and meaning of material components of megalithic settings.
Standard practice, however, necessitates the separation of archaeological materials as defined by modern, western science to facilitate particular forms of analysis before a process of documentary reassembly is undertaken (Conneller 2011, pp. 125, Lucas 2012. The severing of material connections prior to analysis risks precluding the possibility of understanding materials as contextually situated assemblages, meaningful in their composition, their qualities emerging through and specific to their relations. The depositional assemblage is therefore key and provides a baseline for analysis.

LEAVING NO STONE UNTURNED: A RE-EVALUATION OF THE AVEBURY STONE SETTINGS
Excavation by Keiller (Smith 1965) and subsequent work by Gillings et al. (2008) reveal that the Avebury stone settings incorporate cut features of varying dimensions that would have contained megalithic sarsen uprights. Composition of fills varies: as previously noted, many contain small sarsen boulders and blocks, typically 40-50 in the settings within the henge (Smith 1965, p. 222), and a clean, chalk-free, possibly riverine clay, both in direct contact with the sarsen megalith where still extant (Fig. 4). The presence of chalky rubble is recorded in some settings and large chalk blocks in others; stake holes and packed chalk or sarsen surfaces interpreted as ramps used in the erection of the megalith are also noted in a number of examples (Smith 1965, p. 219). Other forms of artefactual material more commonly associated with consumption and occupation debris are remarkable for their near absence (Gillings et al. 2008, pp. 119-120;Gillings and Pollard 2016), and where present, Smith (1965, p. 226) infers accidental inclusion. From this evidence, some key themes emerge: the direct physical connection between the megalithic, clay and small sarsen elements; absences, in the form of the sterility of the clay and the apparent exclusion of particular materials in the stone holes; and variance in composition of assembled components. The importance of deliberate choice of material is highlighted in all three themes, and connections are revealed through exploration of material sources and citation (Jones 2012), through which particular, ontologically conversant qualities are foregrounded.

THE RIVER RUNS THROUGH
The nature of the clay deposits in the stone settings, notably its cleanness, was suggestive to Smith of a riverine source (1965, p. 221), 1 and, deposited as it is, in association with small sarsen blocks and boulders, recalls the structure of a river bed as it moves along its course (cf. Richards [1996aRichards [ , 1996b for a discussion of structural references to the wider landscape in Neolithic monumental architecture). The qualities of clay that emerge within such an environment are those of impermeability, plasticity and adhesiveness, acting to restrict and control the movement of the water (Schumm 2005, pp. 97-103), but when taken from that context and exposed to running water, small quantities of clay can be observed to react to the force and comingle with the liquid, changing its material properties and its efficacy as a medium of containment. Stone also affords stability in a riverine setting, spreads of rocks, boulders and pebbles creating a lining for the river bed, but if removed and redeposited elsewhere will allow water to seep through. It is through the association and interaction between materials that the river bed achieves resilience, providing a flexible, durable and adaptive container to direct moving bodies of waterwater that may quickly disappear if it moves beyond the boundaries of its channel to permeable bedrock. The material qualities described are relational and it is the collaboration between different elements that is therefore key: river clay is a fixative, a quality reinforced by its stickiness, an adhesive capable of binding different materials within a single matrix; the stones lining the river bed offer resistance to the force of the moving water; and what is generated through the combination of clay and stone is a powerful substance of containment.
This line of argument raises a number of questions: are these qualities transferrable and can they make sense in the context of a megalithic setting? If the clay-with-stone is a substance of containment and restraint, what is it that requires this form of negotiation? The relationship between the clay and small sarsen boulders and the megalithic sarsens is next considered, extending and reinforcing the riverine connection, a connection that evokes other potential associations especially with henge monuments, as Colin Richards (1996a) has suggested. Sarsen is a Tertiary sandstone formed in marine sediments (Bowen and Smith 1977, p. 187, Geddes 2000, pp. 60-64, Lawson 2007; the megalithic slabs have gently undulating surfaces and often feature narrow tubular voids produced by the former presence of plants, creating a visual reference to watery environments (Fig. 5). Sourced as individual blocks that typically form large swathes over the downs and along valleys, they have drawn direct comparison with rivers (Long 1858, p. 30). Avebury's megalithic sarsens are each positioned within a discrete cutting and, where present, clay and sarsen boulders are placed in direct contact with the base of the stone. Contained below the line of the surface, their setting suggests a foundational role. This relationship places further emphasis upon the fixative characteristics of the clay and sarsen boulders, and simultaneously reveals in the megalithic sarsen dynamic, vital, mercurial qualities that necessitate containment, qualities associated both with moving bodies of water and brought forth through the mechanical demands of maintaining massive, rigid sarsen blocks in an upright position. Evidence that the megalithic sarsens did indeed prove thus animate is provided by Keiller's stone 14 (Smith's no. 33). In the unpublished account of the excavations of the north-west sector, Keiller describes the stone and its setting in great detail, arguing that its apparent misalignment with the other stones of the outer circle is the result of unintended movement during the process of erection. The sub-surface elements afford the means by which he develops this hypothesis: compression of areas of the chalk bedrock forming the base and sides of the setting indicative of slippage; the small sarsen embedded in the side of the wall of the setting, the result of intense force; and the presence of a void that had been fashioned to receive the base of the megalith filled instead with clay and small sarsens. That no attempt appears to have been made to adjust the stone is noted by Keiller, although his explanation that the position was accepted as 'adequately satisfactory' (Keiller 1937) seems discordant with the care which attended selection and arrangement of associated materials comprising the settings. Approached as an animate, agentive material, however, the decision to secure the megalithic sarsen as found can be understood as a negotiated compromise, whereby the stone is anchored, albeit in a position of its choosing.
Gillings and Pollard (forthcoming) examine the transport and erection of the Avebury megaliths and also suggest that their treatment may reveal concern with a propensity for movement, specifically for return to their respective sources. The character of the fills in the stone settings is briefly noted and contrasted with the rich deposits of animal bone and ceramic sherds identified in features in the West Kennet Palisade enclosures, which, they argue, may be sites of stone extraction. Although the absence of such material in the stone settings is not further explored, discussion focusing on the relationship between the stones, extraction sites and associated deposits, which are interpreted as evidence of feasting and conceived of in terms of meaningful material transactions, this difference is striking and requires explanation. An absence remains a relationship, albeit negative (cf. Fowles 2010), and in support of the proposition developed in the current study may be argued to replicate the relative artefactual sterility of an active river channel, thereby reinforcing the potency of the fixative. The contextually contingent, relational qualities of the clay and sarsen assemblages, specific in their composition, challenge the existence of universal material types each with a suite of fixed, essential attributes.

MATERIAL CLASSES: COMING UNSTUCK
What emerges is a unique material taxonomy, revealing of particular ways of being: the Avebury megalithic settings are an act of disclosure. Sarsen is not a single type, differentially deployed to undertake two counteractive roles of stabiliser and stabilised, but rather two separate materials (cf. Conneller 2011) with potential for different forms of animacy; one anchoring, the other vital and mercurial. The distinction is defined through association, form and behaviour; the small sarsen blocks and rounded boulders show visual affinity with river stone and are found interspersed in 'rivers' of sarsen slabs, the form of which recalls and presences the river itself. The physical act of stone erection draws forth the potential for animacy, the megalithic element either compliant or resistant to its setting and given to sudden, urgent, irresistible movement; and the boulders and small stones mutually conversant, seeking their own stability through settling alongside the megalith within the cutting, in combination with and giving durability to the adhesive clay. There are also at least two clays. Through its active selection and transportation, the clean clay that references riverine deposits is defined as different from clay with inclusions, a point that will be further explored. Its qualities are particular, contextual and transferrable, and its combination with sarsen boulders confirms that these attributes are relational.
One aspect of the settings that has yet to be considered is the differences observed in their material composition. Where present, there is significant variance in the numbers of boulders included: does this inform on different potentials for individual megaliths to move or, rather, varied levels in the potency of the boulders and small sarsens with clay to affect fixity? What of the noted chalk deposits? Sections indicate that clay with sarsens was, at least in some instances, sealed by a fine chalk rubble (Smith 1965, p. 220). The abutment of these two materials is suggestive of deliberate juxtaposition, the powdery aridity of the chalk adhering to the damp stickiness of the clay, both creating a defined, sealed surface and heightening the contrasting qualities of the two materials. However, also present in some settings of the south-east sector are substantial chalk blocks (Smith 1965, p. 221). Interpreted by Smith as packing material, albeit a 'major deviation from the rule' (1965, p. 221) and identified for its potential value for informing on the chronological relationship between the bank and stone settings, the reason for its use goes unexplored. As compact, massive volumes, these chalky elements relate quite differently to the others of the settings, in comparison with the chalk rubble. Do we therefore also have two chalks? Can the chalk blocks be understood in some way as analogous with the small sarsens?
Clearly, exploration of the relational assemblages of the settings raises as many questions as it answers. However, it does highlight similarity, difference and disjuncture both within and between individual settings, providing an argument that material components of each are negotiated as distinctive groupings and that qualities are drawn out through connections that can be contextually cross-referenced. It also places emphasis upon the significance of sub-surface features, which are particular, meaningful and inseparable from the megalithic element manifest at and above the ground surface, and given the widely documented practice of pit deposition in the Neolithic (Thomas 1991, pp. 59-64, 1999, pp. 64-74, 2013, Bradley 2007) this should be unsurprising.

COHESIVE HOLES: SMALL PIT DEPOSITS ON THE WEST KENNET AVENUE OCCUPATION SITE
The West Kennet Avenue occupation site is intimately connected with the stone settings of the Avebury complex, intersecting the course of the West Kennet Avenue and seemingly respected by the absence of a megalithic setting. Excavated by Keiller in 1934 and by Pollard and Gillings over three seasons from 2013 to 2015, it is marked by dense spreads of worked lithics and pot sherds, stake holes and cut features, including a series of post holes, pits and holes (Smith 1965, pp. 210-216, Pollard et al. 2015. Deposits in the small pits, described by Smith as 'holes', invite comparison with the assemblages of the stone settings. They contain occupation debris including worked flint, pottery sherds, burnt daub and charcoal potentially gathered from the surrounding site  and are lined with clay, and sarsen and flint nodules, a material form that cites both the river bed and the stone settings of the henge. However, whereas the clay utilised in the megalithic settings is characterised by its sterility, the clay chosen to line the holes is described as 'dirty' (Smith 1965, p. 212). 2 The use of clay in combination with sarsen and flint suggests an analogous concern with containment in this new context, both referenced by and referencing the qualities foregrounded in the stone settings, and once again implies a vibrancy in the assembled fill. The use of 'dirty' clay is a significant difference, however, and confirms the postulated existence of a second form of the material, perhaps a separate material, appropriate for use in this set of circumstances.
This raises the question of whether the material contained within is of a different character and less prone to animacy than the megaliths and therefore demands a less potent form of containment. The problem with this argument is that it forces a return to a reification of material types as understood by modern science; pure, ideal forms that are subject to contamination through combination with others. Consideration of the source of the material may, once again, prove instructive. The dirty clay may be extracted from a different, non-riverine source, perhaps one associated with an occupation locale, or alternatively may have a known biography that has brought it into contact with other substances. Therefore, through citation of the fixative power of the materials that reference the river bed (Jones 2012), the dirty clay forms a different form of bedding with the capacity to fix animate material assemblages, manifest and expressive of potentially fragile or volatile sets of relations, in the correct place to be remembered or forgotten.
The connections and differences that have emerged between the materials of the stone settings and the small pits, and the two forms of clay in particular, suggest that in the context of the Avebury complex, the placement and fixing of certain materials demanded particular forms of treatment. The need to exclude material associated with occupation from megalithic settings to affect containment and prevent movement, or the necessity to keep different kinds of assemblages of animate, unstable and potentially powerful materials separate, may explain the absent megalithic setting on the West Kennet Avenue.

MEGALITHS, MATERIALITY AND MEANING
The adoption of an assemblage-based approach to explore Avebury's megalithic stone settings has proven insightful, elucidating aspects of the ontologies of those who created them. Emergent through relations and experienced through practice, materials were agentive, with potential for animacy (cf. Gillings and Pollard 2004, pp. 69-70). They were meaningfully deployed, drawing together and presencing different assemblages of materials including people, places, events and actions, so people and other substances could have real effects and make differences in the world. This has implications for the ways in which Avebury can be understood. The sarsen megaliths are revealed as powerful, vital presences with affective qualities that demanded careful, responsive treatment, evidenced by the unique composition of each setting and the anchoring qualities of the sub-surface materials. As a focused gathering of such potent presences, the complex would have had a charged atmosphere, the megaliths constrained yet crackling with dynamic potential, supporting the contention that ongoing contact with the megaliths and movement within the complex could have been restricted or may have required careful negotiation (Gillings and Pollard 2004, pp. 62-70). Whereas Gillings and Pollard (2004, pp. 69-70) draw on ethnographic examples to suggest that the stones may be thought of as ancestors, the observations made in this study argue instead for an understanding that places emphasis on their relationally emergent qualities, the forces of the relations made manifest in, of and through the stones and their settings. This shift in approach makes possible the development of a more nuanced understanding of their nature, and what is revealed is something elusive and nebulous, an indefinable, volatile, imminent presence.
Through the act of gathering, place emerges. The Avebury complex references and draws in elements of the landscape, notably the river and sarsen fields, but in doing so also creates difference: a different assemblage with different affective possibilities that change through ongoing interactions. Concern with the constitutive elements of place is further evidenced through the contrast between the clay used in the megalithic settings and the holes associated with the West Kennet occupation site. The nature of activities and events that were possible at each were discrete, suggesting the existence and acceptance of rules to maintain this distinction: any encounter between materials from the occupation site and those comprising the stone settings would have (potentially undesirable) consequences, creating new assemblages and new substances with new affects. Therefore, although delineated, the Avebury stone settings, the West Kennet occupation site and their constituent assemblages were in a state of responsive tension, prone to sudden change. In tension and intention: these were places of action.
As relationally manifest, every assemblage is unique and it follows that any interpretation must be particular. Nonetheless, an interesting parallel can be seen between the material composition of the Avebury stone settings and that of the Goggleby Stone, part of a megalithic alignment in Shap, Cumbria (Clare 1978). In two episodes of excavation undertaken in 1973 and 1975 in advance of re-erection, it was revealed that the setting included layers of clay and clay with stones. The clay is interpreted in functional terms as a medium to stabilise and enable manipulation of the megalith into a vertical position (Clare 1978, p. 10) and the clay with stones as support and packing. Largely devoid of artefactual material, the exception being a chert chip (Clare 1978, pp. 10-14), the similarity with the Avebury settings at first appears striking. However, the presence of brown patches in the clay thought to represent soil inclusions is an important point of difference (Clare 1978, p. 10).
As a more geographically proximate comparative, the settings of the megalithic elements of Stonehenge provide similarities and some sharp contrasts. Stonehenge has a complex history of construction, deconstruction and reconfiguration. Synthesis and publication of the archaeological investigations undertaken throughout the 20th century, coupled with the findings of more recent programmes of excavation, provide a wealth of detailed information on the composition of the stone settings, whilst documenting the biography-in-progress of the monument in its landscape , Darvill et al. 2012, Parker Pearson 2012. The megalithic element of Stonehenge comprises sarsen blocks thought to have been transported to Salisbury Plain from the sarsen beds near Avebury (Parker Pearson 2012, pp. 292-302, 2015, and also the smaller bluestones, a group comprising different geological types with sources identified, some to outcrop, to the Preseli Hills, Pembrokeshire in Wales (Thorpe et al. 1991, Darvill et al. 2012, Parker Pearson 2012, pp. 261-291, Bevins et al. 2014). Clear differences can be seen between groups of settings. For example, the settings of the Bluestone Circle have a characteristic fill of chalk rubble (Cleal 1995, pp. 224-243) with few additional deposits, typically stone chips (1995, p. 229), whereas the settings of the sarsen circle comprise more heterogeneous fills, including chalk rubble, smaller 'packing' stones of various geological types present in differing proportions, wood ash and earth (1995, pp. 188-206).
Both the Goggleby and Stonehenge settings, approached as assemblages, hold great promise for revealing aspects of past ontologies, of material engagements, with potential to develop interpretations in different directions. This approach also has scope for extension to other megalithic structures, to reconfigure current modes of engagement with, and understanding of, the past.

CONCLUSIONS: RELATIONAL MATERIALITY -A STICKING POINT?
In this study, the megalithic settings of Avebury henge have been examined as diverse material assemblages encompassing elements both above and below the ground surface. This has revealed great complexity, confirming that materials forming the subsurface elements of the assemblages are inseparable from those perceptible above ground; critical structural components resonant with meaning that enable a different understanding of the individual settings and material engagements in the wider Avebury complex. Following Conneller (2011), attention to material qualities emergent through relations exposes connections that can elucidate aspects of past ontologies: two different sarsens, two different clays and two different chalks have revealed themselves, demonstrating that the imposition of material types as defined by modern scientific standards has potential to place limitations on understanding. It has also proven that examination of citation of material relations in different contexts can reveal acts of deliberate variation, key for an exploration of the subtleties of meaning (Jones 2012).
Material qualities are here understood to emerge through relationships, both positive and negative, and are therefore transient, specific but with the capacity for variance at different scales. They are both constitutive of and constituted by context, but also have mobility, and so are transferrable with the potential for cross-contextual reference. This, then, is a call to foreground materials and assemblages, but it comes with a note of caution. Although the value of attendance to materials to inform on past interactions has been demonstrated, the interpretations here tendered should not be taken as universal; they pertain to the examples explored, situated in particular contexts in Neolithic Avebury, and it would be wrong to seek to project the existence of analogous multiple sarsens, clays and chalks onto others. The broader implications of this approach could, however, be drawn upon. It is through alertness to the detail, to the specific configurations of assemblages, that connections and new understandings may be given the opportunity to emerge.
inclusions in the matrix of the 'dirty' clay: 2 and 5 contain flecks of charcoal; 9 contains 'small flint nodules'; and 1 is a yellow colour and contains charcoal and burnt daub.